Minimum Hellinger distance classification for passive sonar in the presence of noise

• Published in: The Journal of the Acoustical Society of America 2010-03, Vol.127 (3_Supplement), p.2043-2043
• Main Authors: Bissinger, Brett E., Culver, R. Lee, Bose, N. K.
• Format: Article
• Language: English
• ISSN: 0001-4966; 1520-8524
• DOI: 10.1121/1.3385385

Passive source classification in the underwater environment is a challenging problem in part because propagation through the space- and time-varying media introduces variability and uncertainty in the signal. Acoustic propagation models can predict received fields accurately, but they are sensitive to input environmental parameters which cannot be known exactly. This uncertainty in environmental knowledge used in signal predictions results in imperfect statistical class models. Classifiers that rely on simulations of the environment must therefore be robust to these imperfections. The minimum Hellinger distance classifier (MHDC) has been shown to be robust to such mismatches in situations with no noise. Low-noise situations allow demonstration of the properties of the classifier; however, real passive sonar tends to have significant noise levels. Therefore the MHDC’s performance is examined and compared to that of a log-likelihood ratio classifier when applied to narrowband passive underwater acoustic signals in the presence of noise. Both classifiers are applied to synthetic Gaussian data, synthetic acoustic data, and actual acoustic data, all with noise. Classifiers are evaluated using receiver operating characteristic curves, a traditional performance metric in signal processing. [Work supported by ONR Undersea Signal Processing.]